Effects of a Population Bottleneck on Whooping Crane Mitochondrial DNA
Variation STOR
®Travis C. Glenn; Wolfgang Stephan; Michael J. Braun
Conservation Biology, Vol. 13, No. 5. (Oct., 1999), pp. 1097-1107.
Stable URL:
http://links.jstor.org/sici?sici=0888-8892%28199910%2913%3A5%3C1097%3AEOAPBO%3E2.0.CO%3B2-H
Conservation Biology is currently published by Blackwell Publishing.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://wvyvy.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/black.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].
Effects of a Population Bottleneck on Whooping Crane Mitochondrial DNA Variation
TRAVIS C. GLENN,*tt WOLFGANG STEPHAN,* AND MICHAEL J. BRAUNf
* Department of Zoology, University of Maryland, College Park, MD 20742, U.S.A.
tLaboratory of Molecular Systematics, MRC 534, Smithsonian Institution, Washington, D.C. 20560, U.S.A.
Abstract: The Whooping Crane (Grus americana^ is an endangered bird that suffered a severe population bot- tleneck; only 14 adults survived in 1938. We assessed the genetic effect of this human-caused bottleneck by se- quencing 314 base pairs (bp) of the mitochondrial DNA control region from cranes that lived before, during, and after this bottleneck. The maximum length of DNA amplifiable from museum specimens was negatively correlated with age, and only 10 of 153 specimens yielded the entire 314 bp sequence. Six haplotypes were present ainong the prebottleneck individuals sequenced, and only one of these persists in the modern popula- tion. The most common modern haplotype ivas in low frequency in the prebottleneck population, which dem- onstrates the powerful effect of genetic drift in changing alíele frequencies in vety small populations. By com- bining all available data, we show that no more than one-third of the prebottleneck haplotypes survived the hmnan-caused population bottleneck. High levels of variation of substitution rates among nucleotide sites prevented us from estimating the prebottleneck population size. Our data ivill be incorporated into the cap- tive breeding program to alloiv better management decisions regarding the preservation of current genetic di- versity. These data offer the first glimpse into the genetic toll this species has paid for human activities.
Efectos de un Cuello de Botella Poblacional en la variación ADN mitocondrial de la Grulla Americana
Resumen: La Grulla Americana (Grus americana^ es un ave amenazada que sufrió un cuello de botella po- blacional severo con tan solo 14 adidtos sobreviviendo en 1938. Evaluamos el efecto genético de este cuello de botella causado por humanos secuenciando 314 pb de la región control del ADN mitocondrial de grullas que vivieron antes, durante y después de este cuello de botella. La longitud máxima de ADN ajnplificable de especttnenes de museo estuvo negativamente correlacionada con la edad, y solo 10 (de 153) especímenes produjeron la secuencia completa con los 314 pb. Cinco haplotipos estuvieron presentes entre los individuos pre-cuello de botella secuenciados y solo uno de estos persiste en las poblaciones modernas. El haplotipo mod- erno tnás común era de baja frecuencia en la población pre-cuello de botella, lo cual demuestra el poderoso efecto de la deriva génica en frecuencias cambiantes de alelos en poblaciones muy pequeñas. Al cojnbinar to- dos los datos a la mano, mostramos que no más de una tercera parte de los haplotipos pre-cuello de botella sobrevivieron a un cuello de botella ocasionado por hutnanos. Los altos niveles de la tasa de variación entre sitios, impidió la estimación del tamaño poblacional previo al cuello de botella. Nuestros datos pueden ser in- corporados al programa de reproducción en cautiverio para perínitir decisiones de manejo mejor enfocadas para la preservación de la diversidad genética actual. Estos datos ofrecen un primer vistazo a el precio genético que esta especie ha pagado debido a actividades humanas.
Introduction
Population bottlenecks, natural or human-induced, can
% Current address: Savannah River Ecology Laboratory, Drawer E, i-,ave significant genetic effects, SUCh as reduced effective Atken, SC 29802, U.S.A., emailglenn@sreLedu , . . ^,• . , ,•.,•^ , ,., • ,•.T • Paper submitted December 30. 1997; revised manuscript accepted population Size (Wngllt 1938), loSS of heteroZygOSlty (Nei January 13, 1999. et al. 1975), and loss of alíeles (Maniyama & Fuerst 1985), 1097
Conservation Biology, Pages 1097-1107 Volume 13, No. 5, October 1999
1098 Whooping Crane mtDNA Variation Glenn et al.
which may lead to decreases in competitiveness, disease resistance, and survival (Soulé 1986; Avise 1994; Keller et al. 1994). Populations that have been reduced in size due to recent human activities have become a focus of conser- vation genetic studies (Bonnell and Selander 1974; Rails et al. 1979; Briscoe et al. 1992; Taylor et al. 1997). Al- though laboratory and manipulative experiments have demonstrated some effects of bottlenecks on controlled populations (Hedrick et al. 1996), studies are needed that directly examine the genetic consequences of bottle- necks on natural populations (y^den et al. 1997).
The Whooping Crane (firus americana) is North Amer- ica's tallest bird and the rarest of the world's 15 crane spe- cies (Johnsgard 1983). Whooping Cranes inhabited much of North America prior to its settlement by Europeans (Fig.
1; Allen 1952). Tlie extent of gene flow among these early populations is unknown. Habitat alterations on the pri- mary breeding grounds by European settlers, combined with hunting for food, sport, and market, led to the inexo- rable depletion of Whooping Crane populations (AUen 1952). All living Whooping Cranes descend from a single population that breeds in Wood Buffalo National Park
(WBNP), Canada, and winters in the Aransas National Wildlife Refuge (ANWR), Texas (Fig. 1; Allen 1952, 1956).
The ANWR Whooping Crane population consisted of only 14 adults in the fall of 1938 and has been carefully moni- tored each year since (Allen 1952; Doughty 1990). Addi- tional Whooping Cranes persisted as a year-round resident population along the Louisiana Gulf coast until 1950 (Doughty 1990), but all of the Louisiana Whooping Cranes and their descendants have subsequently perished. As of February 1997, l60 Whooping Cranes were wintering in the ANWR area, 100 were in captive breeding popula- tions, and 73 were in reintroduced wild populations (Jones & Mirande 1997; Stelin 1997).
Wliooping Cranes offer a unique opportunity to study the loss of genetic variation in a natural population that experienced a severe bottleneck because the census size of the population during its crash and subsequent re- bound have been carefully documented (Allen 1952;
Doughty 1990). Wliooping Cranes have also been the fo- cus of several studies of genetic variation (e.g.. Dessauer et al. 1992; Longmire et al. 1992; Snowbank & Krajewski 1995; Glenn et al. 1997). Because offspring from all sur-
^--^-'^u.C^t
Figure 1. Former primary breeding grounds (shaded area; Allen 1952) and current distribution (named locations) of Whooping Cranes. Whooping Cranes were formerly found throughout much of North America (Allen 1952).
Glenn et al Whooping Crane mtONA Variation 1099
viving maternal lineages were captured for the captive population (Mirande et al. 1991; Jones & Mirande 1997), a complete census of the species' remaining mitochondrial DNA (mtDNA) diversity can be obtained from a small number of individuals (Snowbank & Krajewski 1995). In addition, hundreds of Whooping Cranes have been de- posited in museum collections before, during, and after the bottleneck (Hahn 1954; Glenn 1997). The large size and number of specimens also allow a unique opportu- nity to explore tlie techniques used to recover DNA from museum specimens. We amplified and characterized the most variable portion of mtDNA from museum skins to make direct comparisons of pre- and postbottleneck ge- netic variation in Whooping Cranes (Taylor et al. 1994).
Methods
Tissue Samples
Tissue samples from 153 museum specimens were col- lected (Table 1; Glenn 1997). Most samples, 131, con- sisted of skin snips (usually <5 cm^) with attached feath- ers. One muscle and 11 bone samples were collected from skeletal specimens. From 10 specimens, skin, bone, and foot pad were taken for a comparison of DNA recovery from different tissues. Fresh tissue or blood samples were available from 158 individuals (Table 1;
Glenn 1997). All samples were collected in accordance with the laws regarding endangered species.
Composite nesting areas (CNAs)•territories where nesting through fledging occurs•w^ere established for the entire WBNP population through field studies begun in the 1950s, and captive breeding has emphasized ge- netic representation of all the original CNAs (Mirande et al. 1991). Thanks to this effort, the maternal history of most birds collected after 1966 is known or can be in- ferred (Snowbank & Krajewski 1995; Glenn 1997). A sample of 17 individuals was selected that should contain all possible mitochondrial lineages still persisting within Whooping Cranes. Thus, two groups of Whooping Cranes were examined to assess the effect of the bottleneck: (1) the prebottleneck sample, consisting of 113 museum specimens collected before 1938, and (2) the postbottle- neck sample, consisting of 10 fresh blood samples and 7 museum specimens collected after 1966.
DNA Extraction
We isolated genomic DNA from fresh tissue using one of three protocols: phenol/chloroform/isoamyl alcohol (PCI) extraction and ethanol precipitation (Sambrook et al.
1989), guanidine thiocyanate and silica (Boom et al. 1990), or diatomaceous earth (Carter & Milton 1993). Explicit pro- tocols are available from the Internet (ftp://onyx.si.edu/
protocols/; fue names TSilicaDNA.rtf and MUD- DNA.rtf ).
We used a digestion protocol modified from that of Pääbo et al. (1988) to isolate genomic DNA from a por- tion of skin approximately 10 X 5 mm (about 10 mg) from all museum skins (including attached feather quills). Isolation was followed by PCI extraction. We used a digestion buffer containing 1 mM Ca^^ with no EDTA to improve the efficiency of proteinase K diges- tion (Sambrook et al. 1989). All bone samples were di- gested in the buffer described by Pääbo et al. (1988). Ex- tracts were washed and concentrated with TLE (10 mM Tris pH 8; 0.2 mM EDTA) in Microcon-30 filters (Ami- con, Danvers, Massachusetts, U.S.A.), and the final vol- ume was adjusted to 50-200 (xL. Additional extraction protocols w^ere tested on skin and foot pad tissues by di- gestion in buffer from Pääbo et al. (1988), our modified buffer, and our modified buffer without coUagenase.
Each extraction was foUow^ed by PCI extraction and con- centration as above. We also used an SDS-Urea extrac- tion (Densmore & Wliite 1991) and QiaAmp Tissue Preps (Qiagen Inc., Chatsworth, California, U.S.A.) ac- cording to the manufacturer's mouse-tail protocol. (Ex- plicit protocols may be obtained from the same server referenced for the fresh tissue procedure under the file name 01dDNA.rtf.)
Contamination Control
Contamination control strategies generally followed Thomas (1994; but see Glenn 1997). All supplies used for museum specimens were purchased new and kept in lab- oratories that had not contained bird DNA or polymerase chain reaction (PCR) amplicons. All pre-PCR manipula- tions were conducted in a dedicated pre-PCR laboratory once species-specific primers had been obtained. The DNA was isolated from museum samples in a second ded- icated pre-PCR laboratory that never contained fresh bird DNA or tissue.
To control carry-over contamination, all PCR reactions contained dUTP rather than TTP (Longo et al. 1990).
The exceptions to this strategy were (1) the initial ampli- fication, cloning, and sequencing of the complete con- trol region, which was completed by late 1994 (before the pre-PCR lab was completed), and (2) final amplifica- tion efficiency tests using TTP and dUTP, which were performed after all sequences had been obtained. Five additional precautions were taken to prevent carry-over contamination. First, the amplification, cloning, and se- quencing of the complete control region for primer de- sign used a single individual (USNM 542868) with haplo- type 1. This individual was used as a positive control for all PCR experiments. Second, all museum skin samples were extracted once before PCR of individuals other than the positive control was done. Third, the primers designed to amplify the 314 base pair (bp) fragment were specific to Whooping Cranes under the conditions
Conservation Biology Volume 13, No. 5, October 1999
1100 whooping Crane mtDNA Variation Glenn et al.
employed in this study (DNA from all other crane spe- cies and a variety of other birds and mammals w^ere tested). Fourth, all initial extractions were amplified for four of the fragments before any were sequenced, so no amplicons containing TTP (from the sequencing reac- tions) existed before most PCR amplifications were
completed. Finally, all samples were examined for con- gruence of sequences from all amplified fragments.
Thus, all haplotypes other than the positive control (haplotype 1), could be attributed absolutely to DNA de- rived from the extractions, not re-amplification of previ- ous PCR amplicons, clones, or sequencing reactions.
Table 1. Whooping Crane samples that yielded at least 250 bp of mtDNA sequence.
Studbook ID Museum ID" Tissue Locality'^ Year'' Haplotype''
Prebottleneck AMNH 354211 Skin • • 2 or 6"'^
HMCZ 042510 skin South Dakota 1893 3
USNM 019171 bone Texas 1893 3
USNM 212985 skin Kansas 1910 3
FMNH 401920 skin Nebraska 1922 3
USNM 371331 skin Nebraska • 3
USNM 273926 skin Nebraska 1923 5
USNM 035445 foot • • 6
FMNH 405302 skin Illinois 1891 7
USNM 288519 bone Kansas 1923 7
USNM 007333 bone NWT 1860s 8
JFBM 017948 skin • • CfJ
Postbottleneck
1020 blood WBNP 1985
1027 blood WBNP 1992
1041 blood WBNP 1992
1042 blood WBNP 1985
1048 USNM 567748 skin WBNP 1971
1054 USNM 567747 skin WBNP 1971
1055 USNM 567744 skin WBNP 1974
1057 USNM 567745 skin WBNP 1974
1059 USNM 567743 skin WBNP 1974
1060 USNM 567746 skin WBNP 1974
1062 blood WBNP 1985
1100 blood WBNP 1992
1032 blood WBNP 1985 2
1047 USNM 567749 skin WBNP 1971 2
1031 blood WBNP 1992 3
1128 blood WBNP 1992 3
1195 blood WBNP 1992 3
Additional samples
• USNM 395058 skin ANWR 1948 \f
1002 USNM 599623 skin ANWR 1949" \f
• USNM 428576 skin ANWR 1951 \f
1019 USNM 601126 wing/skin ANWR 1970 /
• USNM 481801 skin Kansas 1965 /
1038 USNM 532714 skin WBNP 1969 /
• USNM 542868 frozen muscle WBNP • /
1011 USNM 428124 bone Louisiana 1950 i'-'
• USNM 419881 skin ANWR 1951 3"
1018 USNM 480442 skin Louisiana 1964 y-^
• USNM 491260 bone Louisiana 1970 5'J
• USNM 420012 skin ANWR 1951 4"
"AMNH, American Museum of Natural History; HMCZ, Harvard Museum of Comparative Zoology; USNM, United States National Museum (Smith- sonian Institution); FMNH, Field Museum of Natural History; JFBM, James Ford Bell Museum of Natural History (University of Minnesota).
''NWT, Northwest Territories, Canada; WBNP, Wood Buffalo National Park (locality for all birds descended from the WBNP/ANWR population unless specifically collected on the ANWR); ANWR, Aransas National Wildlife Refuge.
' Year the bird, blood, or wing was collected.
''See Table 2 for haplotype information (keyed to number provided here).
'Incomplete sequence (amplicons C, K, I), fExcluded from analysis,
^Incomplete sequence (amplicons C, J, I, K).
'' Year brought into captivity.
'Mac, the last wild Whooping Crane from Louisiana; died at ANWR.
^Captive offspring of Josephine.
Glennetal. Whooping Crane mtDNA Variation 1101
DNA Amplification and Sequencing
An alignment of the complete control-region sequence of one Whooping Crane and four other crane species (Glenn 1997) was used to design seven primer pairs for use in amplification and sequencing attempts (Fig. 2).
Hot Start PCR using Hot Beads (Lumitek, Salt Lake City, Utah, U.S.A.) and liigWy stringent PCR conditions was performed in Perkin-Elmer/Cetus 480 thermocyclers equipped with Cycler Mate heated lids (BioLogic Engi- neering, Inc., Shelton, Connecticut, U.S.A.). The PCR was carried out in 50-JJLL volumes of 50 mM KCl, 10 mM Tris-HCl pH 9, 1% Triton X-100, 1.5 mM MgCl2, 150 JJLM
of each dNTP, 250 jjig/mL ESA (Fraction V, Sigma, St.
Louis, Missouri, U.S.A.), and 0.5 JJLM of each primer with 1 unit Taq DNA polymerase (Promega, Madison, Wiscon- sin, U.S.A.) and 50 ng of modern DNA or 2.5 JJLL of mu- seum DNA extract. The highest annealing temperature (55° for amplicons A, B, C, I, and K; 60° for amplicons J and L), lowest Mg^^ concentration, and lowest number of cycles (25-30 for modem DNA; 30-35 for museum DNA) yielding consistent amplification were employed.
Controls with uracil DNA glycosylase (UDG) included liq- uid MgCl2 rather than Hot Beads, 1 unit of UDG, 30 min- utes at 37° and 10 minutes at 94° before cycling. The PCR products were examined by electrophoresis through 2%
agarose gels containing 1X TBE biiffer and ethidium bro- mide (Sambrook et al. 1989).
Inhibition of PCR by museum extracts was tested by adding 50 pg of modem Whooping Crane DNA to a PCR of each DNA extraction and attempting to amplify ampli-
con B (Fig. 2; Glenn 1997). Only DNA extractions that did not previously amplify this fragment were assayed. Ampli- fication products were electrophoresed and scored visu- ally relative to controls on 2% agarose minigels stained with ethidium bromide.
The PCR products > 150 bp were purified by PEG pre- cipitation (Applied Biosystems 1994), whereas those
<150 bp were purified by agarose gel electi-ophoresis followed by spin filtration (Glenn & Glenn 1994). Se- quences were determined directly from PCR products on an ABI 373 stretch automated sequencer (Applied Biosystems 1994) using Taq FS dye terminator chemistry with 10% dimethyl sulfoxide added.
Data Analyses
Analyses were performed to determine if the pattern of genetic variation in the extant postbottleneck popula- tion is different from that of the prebottleneck sample.
We used Tajima's test (Tajima 1989«) to analyze pat- terns of nucleotide diversity, and the routine Proc Freq of SAS (SAS Institute, Cary, North Carolina, U.S.A.) to as- sess the significance of observed haplotype frequency differences. We investigated relationships among the haplotypes using parsimony analysis and median net- works. For parsimony analysis we used the branch and bound search of PAUP* (Swofford 1997), with all charac- ters treated as unordered and equally weighted. We con- structed median networks for the haplotypes according to Bandelt et al. (1995).
LND6 1 L78 L143 L275 L341
ND6 t-Glu
H127 H230 H393 C-48bp
B-249bp
A-51bp
I-314bp J-151bp
K-117bp L-86bp
3 I
Figure 2. Primers and amplicons used to survey variation in the mtDNA control region of Whooping Cranes. Numbering begins with the first base 3 ' of tRNA-Glu. Primer
numbers correspond to the 3 ' base of primers. Length of amplicons ex- cludes primers. Primer sequences are (5' to 3 '): L-ND6, CCCATA ATA CGG CGA AGGATTAGA; L-78, TAY 1125 ATG CCA CATAATACA TTA CAC
• TA; H-127, CAT GCA CTG GTA TGT t-phe GTC TCT TG; L-143, ATCAAT GCA _^ AGA GAC ACATAC CAG; H-230,
HPhe CCG TAT ATT TTG AGG GAG TWG T; L-275, GCA GTG CCT TAG AAC AAA CTA TGA; L-341, GTC TCT CGG ACCAGG TTA TTTATT; H-393, GAA AGA ATG GTC CTG AAG CTA GTA A; H-Phe, TGC TTT GTG GGT TAA GCTA. LND-6 and H-Phe provided by T Parsons.
(W =AorT, Y = CorT)
Conseivation Biology Volume 13, No. 5, October 1999
1102 Whooping Crane mtDNA Variation Glenn et al.
Results
The complete control region of Whooping Cranes is 1125 bp. Comparing sequences among the crane spe- cies sequenced revealed substantial variation. As ex- pected, the 5' third and 3' third contained substantially more variation than the middle third. The most interspe- cific variation occurred in the first third of the control region, which included an indel of about 50 bp. Thus, we focused our efforts on the 314 bp section between positions 78 and 393 of the Whooping Crane mtDNA control region (Fig. 2; Glenn 1997; Genbank accessions AF114338-AF114377; Genbank numbers for the complete mtDNA sequences were assigned as follows; Siberian Crane [Grus (formerly Bugeranus) Bugeraniis leucogera- nus], AF112371; Blue or Stanley Crane [Anthropoidespar- adíseo], API 12372; and Whooping Crane, AFl 12373).
Degradation of Museum DNA
Among museum specimens with precisely known ages, 32 of 35 specimens <80 years old yielded amplicons B or I (249 or 314 bp; Fig. 2), whereas only one of 58 specimens
>80 years old yielded amplicons of these lengths (Fig. 3).
Thus, there was a substantial decrease in PCR ampliflabiUty of long DNA fragments from specimens of increasing age.
We tested several possible explanations for the de- crease in long ampUcons. First, tests for inhibitory effects of the extracts on PCR demonstrated that only 14 of 142 museum skin or muscle extractions showed substantial PCR inhibition. Only 5 of the 14 completely inhibited PCR under the tested conditions; of these, only 1 was
>80 years old. Thus, increasing inhibition among older samples was not responsible for the observed correlation.
We then chose for further study a representative sub- sample of 10 museum specimens spanning the age of
a. 350 a
300
a
c o o
Q. 150
F
CO 100
E 3
E 0
X
.Mf •*+ * ***
0 20 40 60 80 100 120 140 160 180 200
Age of museum skin (years)
Figure 3- Maximum amplicon size versus age for Whooping Crane museum specimens. Only DNA ex- tractions derived from skins with no PCR inhibition are shown. The 164 bp amplicon derives from primers L438 and H603 (Glenn 1997). A best-fit simple linear decay curve of y = 3-625 • 2.716^ yields r^ = 0.715.
the specimens investigated (31-140 years old). Four DNA extraction protocols were tested on skin samples from these birds. No significant correlation of PCR am- plification and DNA extraction method was found (Glenn 1997). Our standard method (digestion with col- lagenase and CaCl2 in the buffer) did, however, give the highest proportion of extracts supporting synthesis of small fragments (amplicons A). We also tested two addi- tional types of tissue, foot pad and bone, which have been reported to be good sources of ancient and mu- seum DNA (Cooper et al. 1992; Mundy et al. 1996).
Again, we found no significant difference in the recov- ery of amplifiable DNA (Glenn 1997). Replacement of dUTP with TTP in our PCR protocol also had no effect (Glenn 1997).
Finally, we tested the efficiency of primer pairs. Al- though the primer pairs differed in PCR efficiency, no consistent trend with amplicon size was found. Dilution series tests with Whooping Crane genomic DNA derived from frozen tissue indicated that the amount of template needed for consistent amplification varied from 5 pg (amplicons A, B, and C) to 50 pg (amplicons I and K) to 500 pg (amplicons J and L).
Effects on Haplotype and Nucleotide Diversity
We identified tliree haplotypes, 1, 2, and 3, with four polymorphic sites in the 314 bp region sequenced in the sample of 17 postbottleneck individuals, which represent all surviving lineages (Table 2). One additional haplotype that is now extinct, 4, was discovered among seven mu- seum specimens salvaged during 1948-1965. Only 10 of the 113 prebottleneck museum specimens yielded data for the entire 314 bp of sequence (Table 2). There were five haplotypes, 3 and 5-8, with four polymorphic sites among these 10 prebottleneck samples (Table 2). One of two additional prebottleneck samples that yielded 270 bp of sequence represented another unique haplotype, 9. If we assume that all postbottleneck haplotypes were present in the prebottleneck population (although some were undetected in the prebottleneck sample), then there were at least nine prebottleneck haplotypes with eight polymorphic sites (Table 2). Thus, it is clear that the number of Whooping Crane haplotypes has decreased dramatically, dropping from at least nine before the bot- tleneck to three in the surviving population.
In contrast, the bottleneck has had little effect on the nucleotide diversity of Whooping Crane mtDNA. We ex- amined two standard measures of average nucleotide di- versity: n, the average number of pairwise nucleotide differences among haplotypes per bp (Nei 1987), and 6 , which is based on the observed number of segregat- ing sites (Watterson 1975; Table 2). Within the prebot- tleneck sample, n is nearly equal to 6. Thus, Tajima's test statistic (D), the standardized difference between ft and 0 (Tajima 1989«), is close to zero.
Glenn et al. Whooping Crane mtDM Variation 1103
Table 2. Mitochondrial DNA haplotypes detected among Whooping Crane samples.
Variable sites 1122223
90701374 Prebottleneck Postbottleneck Haplotype 40703608 frequency" frequency"
1 TCTGATCC'' 0 12 (61)
2 CCCAATCC'' 0"^ [0.5]'^ 2(14)
3 TCCAATCT 5 3(21)
4 CTCAGTTC 0
(f m"
5 TCCAATTT 1 0
6 CCCAACCC^ V[l,5f 0
7 CCCAATTT 2 0
8 CCCAATCT 1 0
9 CTCAG--T 0'' [1]" 0
Total 10 [\2\f 17 (96)
Variable sites 4[6V ^i^-)
ê
0.0045 0.0038ft 0.0045 0.0045
Tajima's D -0.0379 0.5913
"Numbers in parentheses are living captives that descend from com- posite nesting areas with offspring of known haplotypes.
''Position 348 was unambiguously T for both strands of haplotypes 3, 5, 7, 8, and 9. For haplotypes 1, 2, 4, and 6, this site usually gave a stronger C peak on the light strand and a stronger Tpeak on the heavy strand. This pattern was consistent; only three individuals yielded a stronger C peak for both strands.
'^An additional partial sequence from AMNH 354211 that could be haplotype 2 or 6. Excludedfroin calculations of ñ, 9 , and D.
''One individual from museum sample USNM 420012 (see Table 1).
Excluded fro^n calculations o/ ft, 9 , and D because it was only de- tected itnmediately folloiving the bottleneck, in 1951, and does not survive today.
"One individual from JFBM 17948. Excluded from calculations of ft, G, and D because only 270 bp were amplified and sequenced.
f Total of 12 individuals with 6 variable sites if partial sequences (c and e above) are included.
Discussion
Our discovery of mtDNA variation is in contrast to the report of Snowbank and Krajewski (1995), who found no variation among Whooping Crane mtDNA they exam- ined. Although we sequenced only a portion of the re- gion they surveyed -with restriction en2ymes, ^ve discov- ered variation because sequencing is much more powerful than restriction-fragment analysis of most PCR products. In retrospect, haplotype 1 can be distin- guished from haplotypes 2 and 3 'with a restriction en- 2yme (e.g., Sma I) that ^vas not used by Sno^vbank and Krajewski. No restriction enzymes, however, exist to distinguish haplotypes 2 and 3.
Degradation of Museum DNA
The decrease of maximum PCR amplicon size ^th age and the lack of inhibition by the older extracts suggests that DNA is decaying with time in museum specimens or changing in a way that limits its utility as a template for
PCR. We recovered amplifiable DNA genuinely deriving from Whooping Crane skins much less frequently than might have been expected from previous studies of mu- seum specimens (Thomas et al. 1990; EUegren 1991;
Hagelberg et al. 1991; Cooper et al. 1992; Taylor et al.
1994). Previous studies, however, have examined few specimens older than 80 years (n = 0-5 in Thomas et al.
1990; EUegren 1991; Hagelberg et al. 1991; Cooper et al.
1992; Taylor et al. 1994) and have often utilized shorter amplicons, which increases the chance of successful am- plification. Because of the specific sequence of the region studied, primers used for amplicons of about 100 bp were either inefficient for PCR or amplified regions with no variation. Thus, the variable positions in the region under study here could be ampUfied only rarely in amplicons of smaller size. Although it is possible to obtain more tem- plate by simply using more tissue for DNA extraction and by refining the PCR conditions to amplify from less tem- plate, these strategies increase the potential for contami- nation to affect results and increase the damage to valu- able, often irreplaceable, museum specimens. Our results, therefore, suggest that traditional museum preservation and storage techniques yield specimens that serve as via- ble genetic warehouses (Pääbo et al. 1989; Diamond
1990; Graves & Braun 1992) for only a limited time.
Apparent Lack of Population Structure
Full interpretation of our results requires knov^^ledge about the structure of the prebottleneck Wliooping Crane population(s). For instance, the loss of haplotypes could derive entirely from the loss of other populations (i.e., the WBNP population has the same diversity as it did before the bottleneck). Unfortunately, the small size of our prebottleneck sample precludes statistical infer- ences about population structure of the prebottleneck Whooping Crane population. The USNM 007333 is the only prebottleneck Whooping Crane clearly from the WBNP area. Seven other birds, those from South Dakota, Nebraska, Kansas, and Texas, are likely to derive from subpopulations that at least came into contact with birds from WBNP. Because USNM 007333 and USNM 420012 have haplotypes that are now extinct, it is clear that the WBNP/ANWR population lost haplotypes.
Evidence against strong population subdivision comes from two lineages from the nonmigratory Louisiana pop- ulation that were also sequenced for the entire 314 bp of amplicon I. Both of these birds shared the most com- mon prebottleneck haplotype with migratory Whooping Cranes, haplotype 3 (Table 1). In addition, all three hap- lotypes in the modern sample occurred in both current breeding areas of WBNP, Sass and Klewi, indicating that there is no microgeographic structuring of the extant natural population. We therefore assume, for the pur- pose of analysis, that all Whooping Cranes derive from one large, panmictic population, recognizing the influ-
Consei'vation Biology
Volume 13, No. 5, October 1999
1104 Whooping Crane mtDNA Variation Glenn et al.
ence of this assumption on the validity of our conclu- sions.
Effects on Haplotype and Nucleotide Diversity
Tajima's D is often used to test for deviations from selec- tive neutrality, and the low D value obtained is consis- tent with the hypothesis tliat the prebottleneck Wlioop- ing Crane population w^as in an approximate neutral mutation-random drift equilibrium. There are, however, two caveats. First, the estimation of Ä and 6 is based on a mutation model that assumes that multiple substitu- tions do not occur at the same nucleotide site, an as- sumption that appears to be invalidated by our data. Sec- ond, the estimation of ft and 6 from the postbottleneck sample is based on a constructed sample rather than a random sample. Within the postbottleneck sample, ft exceeds 9 ; thus, D is positive due to the loss of rare alíeles, as expected. The D value, however, is not significantly dif- ferent from zero (Tajima 1989ß). The sampling scheme should not significantly bias the estimation of these param- eters because the effective sizes of the CNAs do not vary.
The contrasting patterns of haplotype and nucleotide diversity are consistent "with a recent reduction in popu- lation size and subsequent evolution of the Whooping Crane population by random genetic drift. The differ- ence between these two measures occurs because hap- lotype diversity (i.e., the number of haplotypes in the population) is strongly affected by the elimination of many rare haplotypes, w^hich happens almost instanta- neously after the reduction of population size, whereas nucleotide diversity reñects the less dramatic change in the average differences among the haplotypes (Nei et al. 1975; Maruyama & Fuerst 1985). As an additional con- sequence of losing rare haplotypes, 0, which is esti- mated from the number of segregating sites in the sam- ple, is expected to decline more rapidly than ft, which is estimated as the average number of pairwise differ- ences per nucleotide site between all pairs of haplo- types (Tajima 1989&). Our data are consistent with these theoretical expectations.
Significant Shifts in Haplotype Frequencies
Although our results did not indicate significant differ- ences in nucleotide diversity between pre- and postbottle- neck populations, a direct comparison of the frequencies of individual haplotypes showed a significant effect. Hap- lotype frequencies in the prebottleneck population were estimated from the prebottleneck sample (Table 2). The frequency of Whooping Crane haplotypes among surviv- ing lineages was estimated both by sampling one chick from each founding CNA and by counting the total num- ber of captive individuals believed to descend from each founding CNA. These estimates are in good agreement (Table 2). An exact test was performed on the 4X2
contingency table of the number of individuals observed for the pre- and postbottleneck samples in four haplo- type categories (haplotypes 1, 2, 3, and 5-8; Table 2).
The pre- and postbottleneck frequencies were signifi- cantly different (p < 0.0001).
The most dramatic frequency shift occurred for haplo- type 1. This haplotype was not observed in the prebot- tleneck population but was found in 12 of 17 (71%) founding maternal lineages of the extant population. To test the hypothesis that this frequency shift was due to random drift associated with bottlenecks, we simulated a Wright-Fisher model (Ewens 1979) for a simple bottle- neck scenario with parameters similar to the observed Wliooping Crane bottleneck. Bottlenecks of six genera- tions were simulated. Female population sizes of 35, 18, 7, 10, 13, and 23 were assumed in generations 1-6 re- spectively. These values are half the observed census size of all migratory Wliooping Cranes, beginning in 1916 and resampling every 11 years (approximately the Whooping Crane generation time; Allen 1952; Doughty 1990). All haplotypes were lumped together, except the rare ones. The initial frequency of the rare haplotype was assumed to be 0.1 because, although haplotype 1 was undetected in the prebottleneck sample, a conser- vative estimate of its frequency is the inverse of the sam- ple size (i.e., the same frequency as the singleton class of haplotypes in the prebottleneck sample). Haplotypes in each generation were selected at random, with replace- ment, from the previous generation.
The proportion of bottlenecks with frequency shifts from 0.1 to 0.71 or greater was estimated from 100,000 simulated bottlenecks. Two percent of the simulations from this model produced a frequency shift for the rare haplotype at least as great as the observed frequency shift for haplotype 1. All reasonable changes to the pa- rameters of the simulations gave similar results. Because our survey indicates there were several rare haplotypes with frequencies around 0.1 in the prebottleneck popu- lation (Table 2) and because the probability that any rare haplotype might experience such a frequency shift is the sum of probabilities for all such haplotypes, the sim- ulations suggest that the frequency shift of haplotype 1 is consistent with a simple bottleneck effect (i.e., a re- duction in population size and genetic drift). For a more complicated model with overlapping generations, wliich is more appropriate for Whooping Cranes, the probabil- ity value resulting from simulation is expected to be higher because the effective population size is smaller when generations overlap (Hill 1972).
Heterogeneity of Nucleotide Substitution Rate
It would be desirable to estimate parameters such as the prebottleneck effective population size from our data, but they do not conform to the assumptions of the mod- els used for such estimations, especially the assumption
Gknn et al. Whooping Crane mtDM Variatwn 1105
of single substitutions per site. Our data are consistent with the high level of among-site rate variation that seems to characterize the evolution of mtDNA control- region sequences (Wakeley 1993). For example, the minimum spanning network of the haplotypes (Fig. 4) il- lustrates that multiple shortest paths exist among haplo- types, indicating multiple substitutions at three sites.
Also although only 8 of the 314 sites varied, a parsimony reconstruction of the relationships among the nine hap- lotypes revealed 25 most parsimonious trees of length 11 (i.e., a minimum of 11 substitutions), again suggest- ing that several sites were multiply substituted.
The small number of variable sites prohibited the use of maximum likelihood to evaluate models of rate hetero- geneity, but the parsimony-inferred distribution of substi- tutions per site allowed a gamma-distributed rates model to be tested (Sullivan et al. 1995). If evolutionary rates
are gamma-distributed, the distribution of substitutions per site will foUo^v a negative binomial, w^hereas uniform rates will produce a Poisson distribution. The Whooping Crane data fit a gamma-distributed rates model (a = 0.044) significantly better than a uniform rates model as judged by a likelihood ratio test (df = 1, x^ = 11.93,p <
0.001, Sullivan et al. 1995). This method is expected to underestimate the amount of rate heterogeneity among sites (Wakeley 1993), yet the estimated gamma shape pa- rameter (a = 0.044) indicates a level of rate heterogene- ity among the most extreme ever reported.
Management Implications
Our data allow a direct empirical comparison of genetic variation before and after critical endangerment of a long-lived vertebrate. The recent bottleneck has had a
f j Extant Haplotypes
Extinct Haplotypes Inferred Haplotypes
Figure 4. Median network (Ban- delt et al. 1995) of Whooping Crane haplotypes. Haplotypes are connected by line segments propor- tional to the number of substitu- tions between haplotypes. Positions of inferred substitutions are indi- cated as slashes through the line segments. Haplotypes are drawn proportionately to their frequency
in the prebottleneck sample. Haplo- types not detected in the prebottle- neck sample but observed in post- bottleneck birds (1, 2, and 4) are drawn to the same scale as singly detected haplotypes. Haplotype 9 should be near haplotype 4 but cannot be unambiguously con- nected to the network.
Conservation Biology Volume 13, No. 5, October 1999
1106 Whooping Crane mtDNA Variation Glenn et al.
dramatic effect on the number and frequency of mtDNA haplotypes in the modern Whooping Crane population.
Significant efforts, liowever, are now being made to pre- serve as muclT genetic diversity in this species as possi- ble, and a multi-stage decision matrix is in place for mak- ing captive breeding decisions (Mirande et al. 1991).
The discovery of variation in the mtDNA of modern Whooping Cranes allows the maintenance of mtDNA di- versity to be added to this matrix. More important, this locus is being used to add significant information to the studbook data of Wliooping Cranes so that all individu- als can now be assigned to known maternal lineages.
The discovery of shared haplotypes among the prebot- tleneck resident Louisiana population and the prebottle- neck migratory population supports the idea that these subpopulations exchanged genes, and it supports cur- rent attempts to recreate a nonmigratory population.
The Whooping Crane represents a classic case of species endangerment due to human activities. Its slow but steady recovery stands as a tribute to the long-term com- mitment of individuals, government agencies, and pri- vate organizations to bring this bird back from the brink of extinction (Cannon 1996).
Acknowledgments
We are thankful for the assistance of our colleagues at the U.S. Museum of Natural History, especially D. Albright, P.
Angle, R. Bavis, R. Browning, R. Brumfield, J. Dean, C.
Dove, G. Graves, C. Huddleston, R. Layboume, R. Ojerio, S.
Olson, T. Parsons, J. Sakamoto, G. Satüer, J. Sullivan, D.
Swofford, D. Waller, K. Winker, and E. Zimmer. We grate- fully acknowledge G. Gee, C. Mirande, N. Businga, C. Kra- jewski, S. Snow^bank, S. Jarvi, M. Miller, and K. Jones for samples and information from captive cranes. We also ac- knowledge the museums and thank the curators and col- lection managers who allowed us to sample Whooping Cranes in their research collections and on display: G. Bar- rowclough and P. Sweet (American Museum of Natural History), K. Parks (Carnegie Museum), R. Vasile and M.
Hennen (Chicago Academy of Sciences), S. Lanyon (Field Museum of Natural History), R. Paynter (Harvard Univer- sity Museum of Comparative Zoology), B. Nelson (Hastings Museum), B. Sclimidt (Iowa Valley Community College), F. Sheldon (Louisiana State University), J. Hall (Putnam Museum), G. Schrimper (University of Iowa), R: Zink (University of Minnesota Bell Museum), and T. Labedz (University of Nebraska). This research has been sup- ported by the Smithsonian's Laboratory of Molecular Sys- tematics. National Science Foundation grant DEB-9321604 (to W. S. and T. C. G.) and contract DE-AC09-76SROO-819 between the U.S. Department of Energy (DOE) and Uni- versity of Georgia's Savannah River Ecology Laboratory (DOE contract number DE-FC09-96SR18546).
Literature Cited
Allen, R. P. 1952. The Wliooping Crane. Research report 3- National Audubon Society, New York.
Allen, R. P. 1956. A report on the Wliooping Crane's northern breed- ing grounds. A supplement to research report 3. National Audubon Society, New York.
Applied Biosystems. 1994. 373 DNA sequencing system user's manual.
Part number 903204, revision. A. Perkin Elmer, Foster City, California.
Arden, S. L., D. M. Lambert, A. G. Rodrigo, and I. G. McLean. 1997. The effects of population bottlenecks on multilocus DNA variation in robins. Journal of Heredity 88:179-186.
Avise, J. C. 1994. Molecular markers, natural history and evolution.
Chapman and Hall, New York.
Bandelt, H-J., P. Forster, B. C. Sykes, and M. B. Richards. 1995. Mito- chondrial portraits of human populations using median networks.
Genetics 141:743-753.
Bonnell, M. L., and R. K. Selander. 1974. Elephant seals: genetic varia- tion and near extinction. Science 184:908-910.
Boom, R., C. J. A. Sol, M. M. M. Salimans, C. L. Jansen, P. M. E. Wer- theim-van Dillen, and J. van der Noordaa. 1990. Rapid and simple method for purification of nucleic acids. Journal of Clinical Micro- biology 28:495-503.
Briscoe, D. A., J. M. Malpica, A. Robertson, G. J. Smith, R. Frankham, R. G.
Banks, and J. S. F. Barker. 1992. Rapid loss of genetic variation in large captive populations of Drosophila Oies: implications for the genetic management of captive populations. Conservation Biology 6:416-425.
Cannon, J. R. 1996. Whooping Crane recovery: a case study in public and private cooperation in the conservation of endangered species.
Conservation Biology 10:813-821.
Carter, M. J., and I. D. Milton. 1993. An inexpensive and simple method for DNA purifications on silica particles. Nucleic Acids Re- search 21:1044.
Cooper, A., C. Mourer-Chauvire, G. K. Chambers, A. von Haeseler, A. C. Wilson, and S. Pääbo. 1992. Independent origins of New Zealand moas and kiwis. Proceedings of the National Academy of Sciences, USA 89:8741-8744.
Densmore, L. D., Ill, and P. S. White. 1991. The systematics and evolu- tion of the Crocodylia as suggested by restriction endonuclease analysis of mitochondrial and nuclear ribosomal DNA. Copela 1991:602-615.
Dessauer, H. C, G. F. Gee, and J. S. Rogers. 1992. Allozyme evidence for crane systematics and polymorphisms within populations of Sandhill, Sams, Siberian, and Whooping Cranes. Molecular Phyloge- netics and Evolution 1:279-288.
Diamond, J. M. 1990. Old dead rats are valuable. Nature 347:334-335.
Doughty, R. W. 1990. Return of the Whooping Crane. University of Texas Press, Austin.
Ellegren, H. 1991. DNA typing of museum birds. Nature 354:113.
Ewens, W. J. 1979- Mathematical population genetics. Springer, Berlin.
Glenn, T. C. 1997. Genetic bottlenecks in long-lived vertebrates: mito- chondrial and microsatellite DNA variation in American alligators and Whooping Cranes. Ph.D. dissertation. University of Maryland, College Park.
Glenn, T. C, and S. J. Glenn. 1994. Rapid elutíon of DNA from agarose gels using polyester plug spin inserts (PEPSIs). Trends in Genetics 10:344.
Glenn, T. C, R. S. Ojerio, W. Stephan, and M. J. Braun. 1997. Microsat- ellite DNA loci for genetic studies of cranes. Pages 36-45 in R. P.
Urbanek and D. W. Stahlecker, editors. Proceedings of the 7th North American Crane workshop. North American Crane Working Group, Grand Island, Nebraska.
Graves, G. R., and M. J. Braun. 1992. Museums: storehouses of DNA?
Science 255:1335-1336.
Hagelberg, E., 1. C. Gray, and A. J. Jeffreys. 1991. Identification of the skeletal remains of a murder victim by DNA analysis. Nature 352:
427- 429.
Glenn et al. Crane intDNA Variation 1107
Hahn, P. 1954. Where is that vanished bird? Royal Ontario Museum, University of Toronto, Ontario.
Hedricli, P. W., R. C. Lacy, F. W. Allendorf, and M. E. Soulé. 1996. Dü-ec- tions in conservation biology: comments on Caughley. Conservation Biology 10:1312-1320.
Hill, W. G. 1972. Effective size of populations with overlapping gener- ations. Theoretical Population Biology 3:278-289.
Jolinsgard, P. A. 1983. Cranes of the world. Indiana University Press, Bloomington.
Jones, K. L., and C. Mirande. 1997. The genealogy of the Whooping Crane (Grus americana). International Crane Foundation, Bara- boo, Wisconsin.
Keller, L. F., P. Árcese, J. N. M. Smith, W. M. Hochachlca, and C. C.
Stearns. 1994. Selection against inbred song sparrows during a nat- ural population bottleneclc. Nature 372:356-357.
Longmire, J. L., G. F. Gee, C. L. Hardekopf, and G. A. Marie. 1992. Estab- lislilng paternity in Whooping Cranes (Grus americana) by DNA analysis. Aule 109:522-529.
Longo, M. C, M. S. Berninger, and J. L. Hartley. 1990. Use of uracil DNA glycosylase to control carry-over contamination in poly- merase cliain reactions. Gene 93:125-128.
Maruyaraa, T., and P. A. Fuerst. 1985. Population bottleneclis and non- equilibrium models in population genetics. II. Number of alíeles in a small population that was formed by a recent bottleneck. Genet- ics 111:675-689.
Mirande, C. R., R. Lacy, and U. Seal. 1991. Wliooping Crane (Grus americana') conservation viability assessment. Captive Breeding Specialist Group, Apple Valley, Minnesota.
Mundy, N. I., C. S. Winchell, and D. S. Woodruff 1996. Tandem re- peats and heteroplasy in the mitochondrial DNA control region of the Loggerhead Shrike (Lanius ludovicianus). Journal of Heredity 87:21-26.
Nei, M. 1987. Molecular evolutionary genetics. Equation 10.5. Colum- bia University Press, New York.
Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic variability in populations. Evolution 29:1-10.
Pääbo, S., J. A. Gifford, and A. C. Wilson. 1988. Mitochondrial DNA sequences from a 7000-year old brain. Nucleic Acids Research 16:
9775-9787.
Pääbo, S., R. G. Higuchi, and A. C. Wilson. 1989. Ancient DNA and polymerase chain reaction: the emerging field of molecular archae- ology. Journal of Biological Chemistry 264:9709-9712.
Rails, K., K. Brugger, and J. Ballou. 1979. Inbreeding and juvenile mor- tality in small populations of ungulates. Science 206:1101-1103.
Sambrook J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual. 2nd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
Snowbank, S. A., and C. Krajewski. 1995. Lack of restriction-site varia- tion in mitochondrial-DNA control region of Whooping Cranes (Grus americana). Auk 112:1045-1049-
Soulé, M. E. editor. 1986. Conservation biology: the science of scarcity and diversity. Sinauer Associates, Sunderland, Massachusetts.
Stehn, T. 1997. 1996/97 winter Whooping Crane survey. Aransas National Wildlife Refuge, Austwell, Texas.
Sullivan, J., K. E. Holsinger, and C. Simon. 1995. Among-site rate variation and phylogenetic analysis of I2S rRNA in sigmodontine rodents.
Molecular Biology and Evolution 12:988-1001.
Swofford, D. 1997. PAUP* 4.0.0d55. Phylogenetic analysis using par- simony (and other methods). Sinauer Associates, Sunderland, Massachusetts.
Tajima, F. 1989«. The effects of change in populations size on DNA polymorphism. Genetics 123:597-601.
Tajima, F. 19890. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123:585-595.
Taylor, A. C, W. B. Sherwin, and R. K. Wayne. 1994. Genetic variation of microsatellite loci in a bottlenecked species: the northern hairy- nosed wombat Lasiorhinus krefftii. Molecular Ecology 3:277-290.
Taylor, A. C, A. Horsup, C. N. Johnson, P. Sunnucks, and B. Sherwin.
1997. Relatedness structure detected by microsatellite analysis and attempted pedigree reconstniction in an endangered marsupial, the northern hairy-nosed wombat, Lasiorhinus krefftii. Molecular Ecology 6:9-19.
Thomas, R. H. 1994. Analysis of DNA from natural history museum col- lections. Pages 311-321 in B. Schierwater, B. Streit, G. P. Wagner, and R. DeSalle editors. Molecular ecology and evolution: ap- proaches and applications.. Birkhauser Verlag, Basel, Switzerland.
Thomas, W. K., S. Pääbo, F. X. Villablanca, and A. C. Wilson. 1990. Spa- tial and temporal continuity of kangaroo rat populations shown by sequencing mitochondrial DNA from museum specimens. Journal of Molecular Evolution 31:101-112.
Wakeley, J. 1993. Substitution rate variation among sites in hypervari- able region I of human mitochondrial DNA. Journal of Molecular Evolution 37:613-623.
Watterson, G. A. 1975. On the number of segregating sites in genetical models without recombination. Theoretical Population Biology 7:
256-276.
Wright, S. 1938. Size of population and breeding structure in relation to evolution. Science 87:430-431.
/-v
^'^'
C>^^.
Conservation Biology Volume 13, No. 5, October 1999
